Skip to content

Scoring

Three outside services rate this repository, and each one measures something different. This page lists every item each rater scores, what it saw, and where the gaps are, so a badge at the top of the README reads down to the line that earned it.

The page is generated by scripts/render_scoring.py from a snapshot of the raters' own data, scoring/snapshot.json; a gate fails the build when the page and the snapshot disagree, and only a deliberate refresh reads the raters again, so every date here is a date something was actually read. Snapshot taken 2026-11-12, UTC.

Rater Result What it measures
OpenSSF Scorecard 9.2 / 10 supply-chain and project security practices, read from the repository by a scanner
OpenSSF Best Practices passing criteria for open source projects, each answered with a justification the site publishes
build-doctrine score 3.5 / 5 how far each rule of build-doctrine is enforced here, from stated to gated and proven

OpenSSF Scorecard

Scorecard v5.6.0 read commit 657d18aab0c6 on 2026-11-12. Each check scores 0 to 10 and the overall score is a risk-weighted average; a check that returns -1 is inconclusive and is left out of the average. The scanner documents every check, with its risk and its scoring, in checks.md.

Check Risk Score What the scanner saw
Binary-Artifacts High 10 no binaries found in the repo
Branch-Protection High 4 branch protection is not maximal on development and all release branches
CI-Tests Low 10 16 out of 16 merged PRs checked by a CI test -- score normalized to 10
CII-Best-Practices Low 5 badge detected: Passing
Code-Review High -1 Found no human activity in the last 16 changesets
Contributors Low 3 project has 1 contributing companies or organizations -- score normalized to 3
Dangerous-Workflow Critical 10 no dangerous workflow patterns detected
Dependency-Update-Tool High 10 update tool detected
Fuzzing Medium 10 project is fuzzed
License Low 10 license file detected
Maintained High 10 30 commit(s) and 9 issue activity found in the last 90 days -- score normalized to 10
Packaging Medium 10 packaging workflow detected
Pinned-Dependencies Medium 10 all dependencies are pinned
SAST Medium 10 SAST tool is run on all commits
Security-Policy Medium 10 security policy file detected
Signed-Releases High 10 2 out of the last 2 releases have a total of 2 signed artifacts.
Token-Permissions High 10 GitHub workflow tokens follow principle of least privilege
Vulnerabilities High 9 1 existing vulnerabilities detected

The checks below ten

What each check reads, and the scanner's own detail lines for the ones that did not reach ten.

  • Branch-Protection, 4: reads the default and release branches refuse unreviewed merges.
    • Info: 'allow deletion' disabled on branch 'main'
    • Info: 'force pushes' disabled on branch 'main'
    • Info: 'branch protection settings apply to administrators' is required to merge on branch 'main'
    • Info: 'stale review dismissal' is required to merge on branch 'main'
    • Warn: required approving review count is 1 on branch 'main'
    • Info: codeowner review is required on branch 'main'
    • Warn: 'last push approval' is disabled on branch 'main'
    • Warn: 'up-to-date branches' is disabled on branch 'main'
    • Info: status check found to merge onto on branch 'main'
    • Info: PRs are required in order to make changes on branch 'main'
  • CII-Best-Practices, 5: reads an OpenSSF Best Practices badge, scored by its level.
  • Code-Review, -1: reads recent changes were reviewed by a person other than their author.
  • Contributors, 3: reads recent contributors from more than one company or organization.
    • Info: found contributions from: manifest-identity
  • Vulnerabilities, 9: reads no open, unfixed vulnerabilities in the project or its packages.
    • Warn: Project is vulnerable to: https://osv.dev/GHSA-g3hv-75wg-m249

OpenSSF Best Practices

The entry is project 14563, at the passing level with 100 percent of that level's criteria met, achieved 2026-09-10 and last edited 2026-11-12. Every answer is a claim the badge holder makes, so each row below carries the justification exactly as the entry states it; the gates and tests named in them are the ones this repository runs.

Basics

Criterion Status Justification
description_good Met The README at https://github.com/manifest-identity/manifest-identity#readme opens by stating what the software does and the problem it addresses.
interact Met Issues and pull requests are open on GitHub; https://github.com/manifest-identity/manifest-identity/blob/main/CONTRIBUTING.md states how to propose changes and report problems.
contribution Met https://github.com/manifest-identity/manifest-identity/blob/main/CONTRIBUTING.md states how a change is proposed and what it must pass.
contribution_requirements Met https://github.com/manifest-identity/manifest-identity/blob/main/CONTRIBUTING.md states the requirements a change must meet.
floss_license Met AGPL-3.0.
floss_license_osi Met AGPL-3.0 is OSI approved.
license_location Met https://github.com/manifest-identity/manifest-identity/blob/main/LICENSE
documentation_basics Met The repository's documentation covers installation and use, and renders as a site at https://manifest-identity.github.io/manifest-identity/.
documentation_interface Met The interface is documented in the repository, the documentation is checked against the running interface, and the application serves interactive interface documentation.
sites_https Met GitHub, HTTPS only.
discussion Met GitHub issues and pull requests.
english Met All documentation and discussion are in English.
maintained Met Actively developed. Every change reaches the default branch through a reviewed pull request with required checks, releases are cut from the maintainer's signed tag and listed on the repository's releases page, and the roadmap in the repository states what comes next.

Change control

Criterion Status Justification
repo_public Met https://github.com/manifest-identity/manifest-identity is public.
repo_track Met Every change is tracked in version control and reaches the default branch through a pull request.
repo_interim Met Every interim change lands as a pull request on the public repository; nothing is batched into releases.
repo_distributed Met Git.
version_unique Met One tag per release, in semantic versioning form; the release process is recorded in the repository's decisions.
version_semver Met Semantic versioning. The current release is the newest tag on the repository's releases page.
version_tags Met Every release is a signed tag; https://github.com/manifest-identity/manifest-identity/tags
release_notes Met Each release at https://github.com/manifest-identity/manifest-identity/releases carries notes, and the repository's roadmap states what comes next.
release_notes_vulns Met SECURITY.md commits to naming every fixed vulnerability in the release notes.

Reporting

Criterion Status Justification
report_process Met GitHub issues, described in https://github.com/manifest-identity/manifest-identity/blob/main/CONTRIBUTING.md
report_tracker Met https://github.com/manifest-identity/manifest-identity/issues
report_responses Met https://github.com/manifest-identity/manifest-identity/blob/main/SECURITY.md states how quickly reports are acknowledged and answered.
enhancement_responses Met Enhancement requests are answered in the issue tracker on the same commitments.
report_archive Met https://github.com/manifest-identity/manifest-identity/issues?q=is%3Aissue
vulnerability_report_process Met https://github.com/manifest-identity/manifest-identity/blob/main/SECURITY.md: private reporting through GitHub security advisories.
vulnerability_report_private Met https://github.com/manifest-identity/manifest-identity/security/advisories/new
vulnerability_report_response Met Within fourteen days per SECURITY.md.

Quality

Criterion Status Justification
build Met The repository documents one way to build and run the software from a fresh clone, with dependencies fixed to exact, verified versions.
build_common_tools Met Common, free tools only; the repository documents the build.
build_floss_tools Met All build tooling is FLOSS.
test Met An automated test suite runs in continuous integration on every change, and a pull request cannot merge until it passes.
test_invocation Met One documented command runs the whole suite; the repository states it.
test_most Met Coverage is held above a floor enforced in continuous integration, and controls are proven by removing each one and requiring the tests to notice.
test_continuous_integration Met Continuous integration runs on every push and pull request, and its checks must pass before a merge to the default branch.
test_policy Met The repository's contributing rules require tests with new functionality, and continuous integration enforces a coverage floor on every change.
tests_are_added Met Every functional change in the history lands with its tests, and continuous integration enforces a coverage floor on every change.
tests_documented_added Met https://github.com/manifest-identity/manifest-identity/blob/main/CONTRIBUTING.md
warnings Met Linting and strict typing run in continuous integration.
warnings_fixed Met The pipeline is clean of warnings; a new warning fails the build.
warnings_strict Met The strictest settings of the linter and the type checker are on.

Security

Criterion Status Justification
know_secure_design Met The repository documents how each request is protected and what the software defends against, and carries a threat model.
know_common_errors Met Input handling is documented and tested: escaping at every exit, formula-safe exports, text-only rendering, parameterized queries.
crypto_published Met Only published algorithms, through libraries; transport security belongs to the deployment platform.
crypto_call Met Cryptography is used only through established libraries and the language's standard library; nothing is implemented in the project.
crypto_floss Met The libraries used are free and open source software.
crypto_keylength Met Library defaults for a published, memory-hard password hash; tokens from the standard library's secure random source at its recommended size.
crypto_working Met No broken cryptographic algorithm is used for any security purpose.
crypto_weaknesses Met No weak algorithms in use.
crypto_pfs N/A The application terminates no TLS itself; transport security is the deployment platform's.
crypto_password_storage Met Passwords are stored only as salted hashes from a published, memory-hard algorithm, through a library.
crypto_random Met Every token comes from the standard library's cryptographically secure random source.
delivery_mitm Met Delivery over HTTPS; releases are signed and carry verifiable provenance.
delivery_unsigned Met Every release asset carries build provenance that a consumer can verify against the platform's log; https://github.com/manifest-identity/manifest-identity/releases
vulnerabilities_fixed_60_days Met None outstanding; dependency advisories are received and reviewed as they arrive.
vulnerabilities_critical_fixed Met None outstanding.
no_leaked_credentials Met The full history is scanned for credentials at every commit and in continuous integration, with findings verified against their providers.

Analysis

Criterion Status Justification
static_analysis Met Static analysis, including security queries, runs on every change and on a schedule, with strict typing and linting; the repository names the analyzers.
static_analysis_common_vulnerabilities Met Security-focused static analysis runs over the code and the pipeline definitions; the repository names the analyzers.
static_analysis_fixed Met No open findings.
static_analysis_often Met Every commit.
dynamic_analysis Met The parsers that read outside input are fuzzed under a memory-error detector on changes and on a schedule, beside property-based tests in the suite.
dynamic_analysis_unsafe N/A Memory-safe language (Python).
dynamic_analysis_enable_assertions Met Assertions are active in every test run.
dynamic_analysis_fixed Met No open findings.

The silver and gold levels are not attempted. Both require more than one maintainer, and silver requires a code of conduct, which D-026 in the doctrine declines for a one-person program; a criterion marked met without its evidence would score zero on the doctrine's own scale, so the entry stops where the evidence stops.

build-doctrine score

The doctrine this repository is built under, build-doctrine, scores each of its rules by how far the repository enforces it, and the badge is the mean over the rules that apply to an application repository. The scorer ran at doctrine commit 67c22b3bd5e8.

Level Name Meaning
0 absent or false the rule is not met, or a claim about it is untrue
1 stated a document says it
2 attested a document says it and names where the evidence is
3 checked on demand a command anyone can run verifies it
4 gated the pipeline refuses a merge that breaks it
5 gated and proven gated, and the repository records a run where the gate fired
Rule Level What the scorer saw
readme 4 README.md present; this scorer runs in CI
license 4 LICENSE present; this scorer runs in CI
security-policy 4 SECURITY.md present; this scorer runs in CI
contributing 4 CONTRIBUTING.md present; this scorer runs in CI
decisions-record 4 101 numbered entries; the README states no count
commit-subjects 4 all 11 recent subjects lead with an identifier; CI walks the messages
pinned-actions 5 all 29 uses are pinned by commit; a workflow audit gates it; proven: https://github.com/manifest-identity/manifest-identity/actions/runs/33565119678
ci-gate 4 12 required checks on the mainline: analyze (actions), analyze (python), application, browser, container, doctrine, links, mutation, page, secrets, workflows, writing
dependency-updates 3 update automation configured
run-instructions 1 README has a run section with a command block
troubleshooting 1 README names the likely failures
counted-figures 4 2 bold figures; a test recounts them from the source
generated-artifact-parity not applicable not applicable to an application repository

Mean over the applicable rules: 3.5. A rule at level one is one the repository states and nothing checks; a rule moves up only when a command, then a gate, then a recorded firing stands behind it, and the scorer never infers the fifth level.

Refreshing this page

python3 scripts/render_scoring.py --fetch   # re-read every rater, then render
python3 scripts/render_scoring.py --check   # what the gate runs

The fetch needs a checkout of build-doctrine beside this repository for the doctrine scorer; everything else is read from the raters' public interfaces. The rendered page and the snapshot commit together, and the gate refuses a page the snapshot does not produce.